WebMining Agents Data Mining Prof Dr Ralf Mller

Web-Mining Agents Data Mining Prof. Dr. Ralf Möller Universität zu Lübeck Institut für Informationssysteme Karsten Martiny (Übungen)

Web-Mining Agents Data Mining Track Uncertainty 2

Uncertainty AIMA Chapter 13

Outline • • Agents Uncertainty Probability Syntax and Semantics Inference Independence and Bayes' Rule Bayesian Networks 4

Agents and environments • The agent function maps from percept histories to actions: [f: P* A] • The agent program runs on the physical architecture to produce f • agent = architecture + program architecture: PC, robotic car, … 5

A simple knowledge-based agent 6

Special case: Information retrieval agents • Agent should be allowed to “extract information” from their host environment … • . . . and “make the information available” to their creator (owner) • Very simple example: Linear regression – Percepts = Tuples from a database table – Extract information = compute a model (e. g. , a line with parameters w 0, w 1) of the data – “Make information available” = send the model (w 0, w 1) to the creator, not the data – Data would be too large anyway in general settings – How can we ensure, data is not pilfered? 7

Agents can use simple machine learning techniques 1. Model class (hypothesis space): 2. Loss function: 3. Optimization procedure: • From the viewpoint of the agent creator, the model g(x | θ*) is seen as information (in particular if all elements of the hypothesis space are equally likely) • For the agent, the model is just data as well 8

Model representation formalisms • Polynomial functions (or polynomial rings, algebraic varieties) • Logics (many kinds) • Probability theory • … 9

Overview • Motivation for probabilistic representations in agent context • Probabilistic representation formalism (Bayesian networks) • “Practical” query answering algorithms • Learning Bayesian network (BN) representations • Apply BN representations for information retrieval 10

Acknowledgement The next slides on the Wumpus game (see the AIMA book) are taken from a course on Artificial Intelligence by Thomas Schwarz, S. J. , COEN, SCU 11

Example: Wumpus world • Performance measure – gold +1000, death -1000 – -1 per step, -10 for using the arrow • Environment – – – – Squares adjacent to Wumpus are smelly Squares adjacent to pit are breezy Glitter iff gold is in the same square Shooting kills Wumpus if you are facing it Shooting uses up the only arrow Grabbing picks up gold if in same square Releasing drops the gold in same square • Actuators: Left turn, Right turn, Forward, Grab, Release, Shoot • Sensors: Stench, Breeze, Glitter, Bump, Scream 12

Excursion: Wumpus world PEAS description • Fully Observable – No – only local perception • Deterministic – Yes – outcomes exactly specified • Episodic – No – sequential at the level of actions • Static Yes – Wumpus and Pits do not move • Discrete – Yes • Single-agent? – Yes – Wumpus is essentially a natural feature 13

Exploring a Wumpus world 14

Exploring a Wumpus world 15

Exploring a Wumpus world 16

Exploring a Wumpus world 17

Exploring a Wumpus world 18

Exploring a Wumpus world 19

Exploring a Wumpus world 20

Exploring a Wumpus world 21

Exploring a Wumpus world • Tight Spot – A pit might be in all new rooms – No safe room • Can use probabilistic reasoning – To do next 22

Exploring a Wumpus world • Another tight spot – Wumpus might be in any of the two new locations • Claim: – With representations for dealing with uncertainty an Wumpus playing agent will play better in the long term – But we need to ensure that • model complexity is “manageable” • typical-case complexity for probabilistic computation problems is “not too high” 23

Uncertainty Let action At = leave for airport t minutes before flight Will At get me there on time? Problems: 1. 2. 3. 4. partial observability (road state, other drivers' plans, etc. ) noisy sensors (traffic reports) uncertainty in action outcomes (flat tire, etc. ) immense complexity of modeling and predicting traffic Hence, it seems that a purely logical approach either 1. 2. risks falsehood: “A 25 will get me there on time”, or leads to conclusions that are too weak for decision making: “A 25 will get me there on time if there's no accident on the bridge and it doesn't rain and my tires remain intact etc. ” (A 1440 might reasonably be said to get me there on time but I'd have to stay overnight in the airport …) 24

Problems of logic in certain domains • Diagnosis: – Toothache => Cavity Gum. Problem Abscess … – Cavity => Toothache • The connection between toothaches and cavities is just not a logical consequence. For medical diagnosis logic it does not seem to be appropriate. 25

Methods for handling uncertainty • Logic: – Assume my car does not have a flat tire – Assume A 25 works unless contradicted by evidence • Issues: What assumptions are reasonable? How to handle contradiction? • Rules with fudge factors (belief in the rule): – A 25 ⊢ 0. 3 get there on time – Sprinkler ⊢ 0. 99 Wet. Grass – Wet. Grass ⊢ 0. 7 Rain • Issues: Problems with combination, e. g. , Sprinkler causes Rain? ? • Probability – Model agent's degree of belief – Given the available evidence, – A 25 will get me there on time with probability 0. 04 26

Probability Probabilistic assertions summarize effects of • laziness: failure to enumerate exceptions, qualifications, etc. • theoretical ignorance: no complete theory • practical ignorance: lack of relevant facts, initial conditions, tests, etc. Subjective probability: • Probabilities relate propositions to agent's own state of knowledge e. g. , P(A 25 | no reported accidents) = 0. 06 These are not assertions about the world Probabilities of propositions change with new evidence: e. g. , P(A 25 | no reported accidents, 5 a. m. ) = 0. 15 27

Making decisions under uncertainty Suppose I believe the following: P(A 25 gets me there on time | …) P(A 90 gets me there on time | …) P(A 120 gets me there on time | …) P(A 1440 gets me there on time | …) = 0. 04 = 0. 70 = 0. 95 = 0. 9999 • Which action to choose? Depends on my preferences for missing flight vs. time spent waiting, etc. – Utility theory is used to represent and infer preferences – Decision theory = probability theory + utility theory 28

Decision theoretic agent 29

Probability theory: Representation formalism • Basic element: random variable • Similar to propositional logic: possible worlds defined by assignment of values to random variables. • Boolean random variables e. g. , Cavity (do I have a cavity? ). Domain is <true, false> • Discrete random variables e. g. , Weather is one of <sunny, rainy, cloudy, snow> • Domain values must be exhaustive and mutually exclusive • Elementary proposition constructed by assignment of a value to a random variable: e. g. , – Weather = sunny, – Cavity = false (abbreviated as cavity) – Cavity = true (abbreviated as cavity) • Complex propositions formed from elementary propositions and standard logical connectives e. g. , Weather = sunny Cavity = false 30

Syntax • Atomic event: A complete specification of the state of the world about which the agent is uncertain E. g. , if the world is described by only two Boolean variables • Cavity and Toothache, then there are 4 distinct atomic events: Cavity = false Toothache = false Cavity = false Toothache = true Cavity = true Toothache = false Cavity = true Toothache = true • Atomic events are mutually exclusive and exhaustive 31

Axioms of probability • For any propositions A, B – 0 ≤ P(A) ≤ 1 – P(true) = 1 and P(false) = 0 – P(A B) = P(A) + P(B) - P(A B) 32

Example world Example: Dentist problem with four variables: Toothache (I have a toothache) Cavity (I have a cavity) Catch (steel probe catches in my tooth) Weather (sunny, rainy, cloudy, snow ) 33

Prior probability • Prior or unconditional probabilities of propositions e. g. , P(Cavity = true) = 0. 1 and P(Weather = sunny) = 0. 72 correspond to belief prior to arrival of any (new) evidence • Probability distribution gives values for all possible assignments: P(Weather) = <0. 72, 0. 1, 0. 08, 0. 1> (normalized, i. e. , sums to 1 because one must be the case) 34

Full joint probability distribution • Joint probability distribution for a set of random variables gives the probability of every atomic event on those random variables P(Weather, Cavity) = a 4 × 2 matrix of values: Weather = Cavity = true Cavity = false • sunny 0. 144 0. 576 rainy 0. 02 0. 08 cloudy 0. 016 0. 064 snow 0. 02 0. 08 Full joint probability distribution: all random variables involved – P(Toothache, Catch, Cavity, Weather) • Every query about a domain can be answered by the full joint distribution 35

Probability for continuous variables 36

Discrete random variables: Notation • Dom(W) = {sunny, rainy, cloudy, snow} and Dom(W) disjoint from domain of other random variables: – Atomic event W=rainy often written as rainy – Example: P(rainy), the random variable W is implicitly defined by the value rainy • Boolean variable C – Atomic event C=true written as c – Atomic event C=false written as c – Examples: P(c) or P( c) 37

P(x) for P(X=x) also used for continuous random vars 38

Conditional probability • Conditional or posterior probabilities e. g. , P(cavity | toothache) = 0. 8 or: <0. 8> i. e. , given that toothache is all I know • (Notation for conditional distributions: P(Cavity | Toothache) = 2 -element vector of 2 -element vectors) • If we know more, e. g. , cavity is also given, then we have P(cavity | toothache, cavity) = 1 • New evidence may be irrelevant, allowing simplification, e. g. , P(cavity | toothache, sunny) = P(cavity | toothache) = 0. 8 • This kind of inference, sanctioned by domain knowledge, is crucial 39

Conditional probability • Definition of conditional probability (in terms of uncond. prob. ): • Product rule gives an alternative formulation ( is commutative): • A general version holds for whole distributions, e. g. , P(a | b) = P(a b) / P(b) if P(b) > 0 P(a b) = P(a | b) P(b) = P(b | a) P(Weather, Cavity) = P(Weather | Cavity) P(Cavity) View as a set of 4 × 2 equations, not matrix mult. (1, 1) P(Weather=sunny |Cavity=true) P(Cavity=true) (1, 2) P(Weather=sunny |Cavity=false) P(Cavity=false), …. • Chain rule is derived by successive application of product rule: P(X 1, …, Xn) = P(X 1, . . . , Xn-1) P(Xn | X 1, . . . , Xn-1) = P(X 1, . . . , Xn-2) P(Xn-1 | X 1, . . . , Xn-2) P(Xn | X 1, . . . , Xn-1) =… = P(Xi | X 1, … , Xi-1) 40

Bayes‘ Rule 41

Inference by enumeration • Start with the joint probability distribution: • For any proposition φ, sum the atomic events where it is true: P(φ) = Σω: ω⊨φ P(ω) 42

Inference by enumeration • Start with the joint probability distribution: • For any proposition φ, sum the atomic events where it is true: P(φ) = Σω: ω╞φ P(ω) • P(toothache) = 0. 108 + 0. 012 + 0. 016 + 0. 064 = 0. 2 • Unconditional or marginal probability of toothache • Process is called marginalization or summing out 43

Marginalization and conditioning • Let Y, Z be sequences of random variables s. th. Y Z denotes all random variables describing the world • Marginalization – P(Y) = Σz ∈ ZP(Y, z) • Conditioning – P(Y) = Σz ∈ ZP(Y|z)P(z) 44

Inference by enumeration • Start with the joint probability distribution: For any proposition φ, sum the atomic events where it is true: P(φ) = Σω: ω⊨φ P(ω) • P(cavity toothache) = 0. 108 + 0. 012 + 0. 072 + 0. 008+ 0. 016 + 0. 064 = 0. 28 (P(cavity toothache) = P(cavity) + P(toothache) – P(cavity toothache)) 45

Inference by enumeration • Start with the joint probability distribution: • Can also compute conditional probabilities: P( cavity | toothache) = P( cavity toothache) Product rule P(toothache) = 0. 016+0. 064 0. 108 + 0. 012 + 0. 016 + 0. 064 = 0. 4 P(cavity | toothache) = 0. 108+0. 012/0. 2 = 0. 6 46

Normalization • Denominator P(z) (or P(toothache) in the example before) can be viewed as a normalization constant α P(Cavity | toothache) = α P(Cavity, toothache) = α [P(Cavity, toothache, catch) + P(Cavity, toothache, catch)] = α [<0. 108, 0. 016> + <0. 012, 0. 064>] = α <0. 12, 0. 08> = <0. 6, 0. 4> General idea: compute distribution on query variable by fixing evidence variables (Toothache) and summing over hidden variables (Catch) 47

Bayes‘ Rule 48

Inference by enumeration, contd. Typically, we are interested in the posterior joint distribution of the query variables Y given specific values e for the evidence variables E (X are all variables of the modeled world) Let the hidden variables be H = X - Y – E then the required summation of joint entries is done by summing out the hidden variables: P(Y | E = e) = αP(Y, E = e) = αΣh. P(Y, E= e, H = h) • The terms in the summation are joint entries because Y, E and H together exhaust the set of random variables (X) • Obvious problems: 1. Worst-case time complexity O(dn) where d is the largest arity and n denotes the number of random variables 2. Space complexity O(dn) to store the joint distribution 3. How to find the numbers for O(dn) entries? 49

Independence • A and B are independent iff P(A|B) = P(A) or P(B|A) = P(B) or P(A, B) = P(A) P(B) P(Toothache, Catch, Cavity, Weather) = P(Toothache, Catch, Cavity) P(Weather) • • 32 entries reduced to 12; for n independent biased coins, O(2 n) →O(n) Absolute independence powerful but rare Dentistry is a large field with hundreds of variables, none of which are independent. What to do? 50

Conditional independence • P(Toothache, Cavity, Catch) has 23 – 1 = 7 independent entries • If I have a cavity, the probability that the probe catches in it doesn't depend on whether I have a toothache: (1) P(catch | toothache, cavity) = P(catch | cavity) • The same independence holds if I haven't got a cavity: (2) P(catch | toothache, cavity) = P(catch | cavity) • Catch is conditionally independent of Toothache given Cavity: P(Catch | Toothache, Cavity) = P(Catch | Cavity) • Equivalent statements: P(Toothache | Catch, Cavity) = P(Toothache | Cavity) P(Toothache, Catch | Cavity) = P(Toothache | Cavity) P(Catch | Cavity) 51

Conditional independence contd. • Write out full joint distribution using chain rule: P(Toothache, Catch, Cavity) = P(Toothache | Catch, Cavity) P(Catch | Cavity) P(Cavity) conditional independence = P(Toothache | Cavity) P(Catch | Cavity) P(Cavity) i. e. , 2 + 1 = 5 independent numbers • In most cases, the use of conditional independence reduces the size of the representation of the joint distribution from exponential in n to linear in n. • Conditional independence is our most basic and robust form of knowledge about uncertain environments. 52